RAS proto-oncogenes are mutated at high frequency in many different malignancies. Thus, developing effective therapeutic strategies for reversing the biochemical consequences of oncogenic Ras is a fundamental obstacle to reducing the worldwide burden of cancer. Although oncogenic RAS alleles encode gain-of-function proteins that are robustly expressed in cancer cells, intrinsic characteristics of the Ras/GTPase activating protein (Ras/GAP) molecular switch pose difficult, if not insurmountable, challenges to developing targeted inhibitors. The undruggable biochemical properties of the oncogenic Ras/GAP switch represent a central unsolved problem in cancer therapeutics. Activated Ras engages a complex network of kinase effector cascades of which the Raf/MEK/ERK and phosphoinositide-3-OH kinase (PI3K), Akt, mammalian Target of Rapamycin (PI3K/Akt/mTOR) pathways are strongly implicated in cancer initiation and maintenance. In this project, we will exploit transplantable primary myeloid and lymphoid leukemias from strains of Nf1 mutant and Kras/Nras knock in mice that accurately model human cancers as a controlled evolutionary system and experimental platform for interrogating responses to small-molecule inhibitors. In particular, we have transplanted ~40 primary leukemias into cohorts of mice, and have treated these recipients with MEK and PI3K inhibitors alone and in combination. We have isolated multiple, independent drug resistant leukemias from these controlled preclinical trials, and have shown that acquired resistance follows distinct evolutionary trajectories. These data recapitulate, with remarkable fidelity, the initial response and ultimate relapse of advanced human cancers treated with targeted inhibitors. This general approach has the additional advantage of providing a tractable forward genetic system for discovering and validating mechanisms of de novo and acquired resistance. Here we propose to use these novel reagents to interrogate in vivo clonal selection of cancers driven by oncogenic Ras signaling in response to treatment with MEK and PI3K inhibitors as well as mechanisms of response and resistance. The specific aims of this PQ proposal are: (1) to investigate the evolution of acquired resistance to MEK inhibitors in primary AML characterized by Nf1 inactivation or by oncogenic Nras/Kras mutations; and (2) to elucidate the clonal architecture, evolution, and drug responses in T-ALLs from wild-type and Kras mutant mice. Our overall goals are: (1) to reveal biologic principles underlying how the selective pressure imposed by MEK and/or PI3K inhibitor treatment leads to clonal evolution of Ras-driven cancers in vivo; (2) to discover specific genes and pathways that confer resistance to targeted anti-cancer agents; and, (3) to use these data to develop therapeutic paradigms for reversing the adverse biochemical outputs of oncogenic Ras that can be translated through human clinical trials.